[PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[Khem Raj]

The fix is to remove a static_assert which is anyway diagnostics
we will not hit on 64bit platforms and it will still work on 32bit
platforms although a bit slow.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
---
 ...move-static_assert-for-64bit-atomics.patch | 33 +++++++++++++++++++
 .../recipes-sato/webkit/webkitgtk-3_2.40.0.bb |  1 +
 meta/recipes-sato/webkit/webkitgtk_2.40.0.bb  |  1 +
 3 files changed, 35 insertions(+)
 create mode 100644 meta/recipes-sato/webkit/files/0001-ANGLE-Remove-static_assert-for-64bit-atomics.patch

diff --git a/meta/recipes-sato/webkit/files/0001-ANGLE-Remove-static_assert-for-64bit-atomics.patch b/meta/recipes-sato/webkit/files/0001-ANGLE-Remove-static_assert-for-64bit-atomics.patch
new file mode 100644
index 0000000000..d39eee6c58
--- /dev/null
+++ b/meta/recipes-sato/webkit/files/0001-ANGLE-Remove-static_assert-for-64bit-atomics.patch
@@ -0,0 +1,33 @@
+From 03d5f84e7bb292f5cf7120af91e979fc87b36e04 Mon Sep 17 00:00:00 2001
+From: Khem Raj <raj.khem@gmail.com>
+Date: Tue, 28 Mar 2023 14:38:17 -0700
+Subject: [PATCH] ANGLE: Remove static_assert for 64bit atomics
+
+This will make sure that it can be built on 32bit platforms e.g.
+mips32/ppc32
+
+This partially reverts https://chromium-review.googlesource.com/c/angle/angle/+/4205892
+the suggestion to remove the assertion is from the author of the patch
+itself
+
+Upstream-Status: Inappropriate [32-bit specific]
+Signed-off-by: Khem Raj <raj.khem@gmail.com>
+---
+ Source/ThirdParty/ANGLE/src/libANGLE/renderer/serial_utils.h | 1 -
+ 1 file changed, 1 deletion(-)
+
+diff --git a/Source/ThirdParty/ANGLE/src/libANGLE/renderer/serial_utils.h b/Source/ThirdParty/ANGLE/src/libANGLE/renderer/serial_utils.h
+index 51ad1af9..9412c805 100644
+--- a/Source/ThirdParty/ANGLE/src/libANGLE/renderer/serial_utils.h
++++ b/Source/ThirdParty/ANGLE/src/libANGLE/renderer/serial_utils.h
+@@ -129,7 +129,6 @@ class AtomicQueueSerial final
+   private:
+     static constexpr uint64_t kInvalid = 0;
+     std::atomic<uint64_t> mValue       = kInvalid;
+-    static_assert(decltype(mValue)::is_always_lock_free, "Must always be lock free");
+ };
+ 
+ // Used as default/initial serial
+-- 
+2.40.0
+
diff --git a/meta/recipes-sato/webkit/webkitgtk-3_2.40.0.bb b/meta/recipes-sato/webkit/webkitgtk-3_2.40.0.bb
index a9d6af2626..23378d7d16 100644
--- a/meta/recipes-sato/webkit/webkitgtk-3_2.40.0.bb
+++ b/meta/recipes-sato/webkit/webkitgtk-3_2.40.0.bb
@@ -15,6 +15,7 @@ SRC_URI = "https://www.webkitgtk.org/releases/webkitgtk-${PV}.tar.xz \
            file://reproducibility.patch \
            file://0d3344e17d258106617b0e6d783d073b188a2548.patch \
            file://disable_wasm_riscv64.patch \
+           file://0001-ANGLE-Remove-static_assert-for-64bit-atomics.patch \
            "
 SRC_URI[sha256sum] = "a4607ea1bf89669e89b1cb2c63faaec513f93de09b6ae60cc71d6a8aab7ab393"
 
diff --git a/meta/recipes-sato/webkit/webkitgtk_2.40.0.bb b/meta/recipes-sato/webkit/webkitgtk_2.40.0.bb
index e219cbf6d4..5c64dab23b 100644
--- a/meta/recipes-sato/webkit/webkitgtk_2.40.0.bb
+++ b/meta/recipes-sato/webkit/webkitgtk_2.40.0.bb
@@ -15,6 +15,7 @@ SRC_URI = "https://www.webkitgtk.org/releases/${BPN}-${PV}.tar.xz \
            file://reproducibility.patch \
            file://0d3344e17d258106617b0e6d783d073b188a2548.patch \
            file://disable_wasm_riscv64.patch \
+           file://0001-ANGLE-Remove-static_assert-for-64bit-atomics.patch \
            "
 SRC_URI[sha256sum] = "a4607ea1bf89669e89b1cb2c63faaec513f93de09b6ae60cc71d6a8aab7ab393"
 
-- 
2.40.0

Re: [OE-core] [PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[Alexander Kanavin]

On Tue, 28 Mar 2023 at 23:43, Khem Raj <raj.khem@gmail.com> wrote:
> +Subject: [PATCH] ANGLE: Remove static_assert for 64bit atomics
> +
> +This will make sure that it can be built on 32bit platforms e.g.
> +mips32/ppc32
> +
> +This partially reverts https://chromium-review.googlesource.com/c/angle/angle/+/4205892
> +the suggestion to remove the assertion is from the author of the patch
> +itself
> +
> +Upstream-Status: Inappropriate [32-bit specific]

32-bit specificity is not in itself inappropriate for upstream
submission, can you please either submit upstream or open a ticket and
refer to it in the patch?

Alex

Re: [OE-core] [PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[Khem Raj]

On Wed, Mar 29, 2023 at 1:38 AM Alexander Kanavin
<alex.kanavin@gmail.com> wrote:
>
> On Tue, 28 Mar 2023 at 23:43, Khem Raj <raj.khem@gmail.com> wrote:
> > +Subject: [PATCH] ANGLE: Remove static_assert for 64bit atomics
> > +
> > +This will make sure that it can be built on 32bit platforms e.g.
> > +mips32/ppc32
> > +
> > +This partially reverts https://chromium-review.googlesource.com/c/angle/angle/+/4205892
> > +the suggestion to remove the assertion is from the author of the patch
> > +itself
> > +
> > +Upstream-Status: Inappropriate [32-bit specific]
>
> 32-bit specificity is not in itself inappropriate for upstream
> submission, can you please either submit upstream or open a ticket and
> refer to it in the patch?

did you read the message trail on the gerrit review I added to commit ?

>
> Alex

Re: [OE-core] [PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[Alexander Kanavin]

On Wed, 29 Mar 2023 at 17:07, Khem Raj <raj.khem@gmail.com> wrote:
> > 32-bit specificity is not in itself inappropriate for upstream
> > submission, can you please either submit upstream or open a ticket and
> > refer to it in the patch?
>
> did you read the message trail on the gerrit review I added to commit ?

I did - upstream suggested a workaround that was submitted here (in an
already closed merge request), but they didn't definitively say
whether they would reject a ticket (or a patch) to solve the problem
properly, so I think it's worth to at least open a ticket.

Alex

Re: [OE-core] [PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[Khem Raj]

[-- Attachment #1.1.1: Type: text/plain, Size: 753 bytes --]

On 3/29/23 12:04 PM, Alexander Kanavin wrote:
> On Wed, 29 Mar 2023 at 17:07, Khem Raj <raj.khem@gmail.com> wrote:
>>> 32-bit specificity is not in itself inappropriate for upstream
>>> submission, can you please either submit upstream or open a ticket and
>>> refer to it in the patch?
>>
>> did you read the message trail on the gerrit review I added to commit ?
> 
> I did - upstream suggested a workaround that was submitted here (in an
> already closed merge request), but they didn't definitively say
> whether they would reject a ticket (or a patch) to solve the problem
> properly, so I think it's worth to at least open a ticket.
> 

that is still under discussion. Hopefully they are amenable for such 
accommodation

> Alex

[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 2613 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 203 bytes --]

Re: [OE-core] [PATCH] webkitgtk: Fix build on 32bit platforms without 64bit atomics

[Alexander Kanavin]

On Wed, 29 Mar 2023 at 21:11, Khem Raj <raj.khem@gmail.com> wrote:
> > I did - upstream suggested a workaround that was submitted here (in an
> > already closed merge request), but they didn't definitively say
> > whether they would reject a ticket (or a patch) to solve the problem
> > properly, so I think it's worth to at least open a ticket.
> >
>
> that is still under discussion. Hopefully they are amenable for such
> accommodation

But then the patch is Pending, and not Inappropriate, and should be
marked as such.

Upstream now suggested it could be submitted, so please do.

Alex